I’m currently reviewing if we can use octopus for our deploys, and I’m trying to work out what (if any) support exists for the scenario where the server with Tentacle installed can’t connect to the one with Octopus installed (We have many environments where we simply can’t get the ports opened).
I was planning to have a manual step defined for those environments on those projects, but where possible I want to actually keep our install process the same to avoid differences in the environments.
I notice tentacle has a deploy-package flag, but I haven’t been able to find any specs on exactly what options are available for this. I also couldn’t find a simple way to actually download the packages for a release. I’m also curious whether there’s any support for the more advanced deploy features like setting up IIS sites & windows services or if this is a connection-only feature currently?
Thanks for getting in touch! When you can’t open ports (and it isn’t behind a proxy) you can use polling tentacles:
We have support coming for offline deployments, where you build the deployment and it’s file from Octopus, that you can then move manually to a server and it will run without the need for a Tentacle.
But this is tied heavily into 3.0 which at it’s most optimistic estimation is 3-4 months away.
Thanks for the reply. I had seen that, but we’re often dealing servers from which we can’t get outbound internet connectivity through the firewalls that are set up either & there’s no real practical way to change that.
I guess that means this isn’t a viable option for us, at least until 3.0 comes out.
For REALLY locked down servers, customers currently either use WIX or install an Octopus Server directly on the locked down server, with its own Tentacle. Then when a project gets to the ‘release to production stage’ they use octo.exe to export from the main OD and import over at the locked down server. This will need to rely on the project not using library variable sets or step templates as these do not export currently.